You’ve probably signed up for a product that no longer exists. You gave it your email, moved your recipes, photos, or workout data into it, and then, one day, a company you’d never heard of bought a company you barely tolerated and quietly turned the lights off. This episode is about that — the “launch fast, iterate, see what sticks” gospel of tech, and the people who get left holding a bin of dead hardware when it doesn’t stick.
Welcome to Episode 9, where Tim and James try to figure out whether shipping fast is genius or just outsourcing your QA department to your paying customers.
The Death of the Viable in MVP
The episode opens on a pet peeve: nobody remembers what the V in MVP stands for. Tim’s had it reframed to him as “most valuable product” and, his personal favorite, “marketing’s viable product” — the landing page with an email signup and absolutely nothing behind it. James pointed out this is now getting worse with AI, since you can spin up a convincing site in an afternoon, frustrate a bunch of people who thought they’d found exactly what they needed, and deliver nothing.
Tim’s actual rule, from years of working with customers, is the one worth writing on a wall: if a customer can do something today, they will expect to do it tomorrow. Which means your “minimum viable product” quietly includes every feature your last version had. See also: launching a music app without shuffle. More on that crime shortly.
The Case for Shipping Garbage (Occasionally)
To be fair, James came armed with the redemption arc. Grand Theft Auto was originally a racing game called Race and Chase, until a glitch made the police cars go feral — ramming players, chasing them down — and the beta testers abandoned the actual race to just get chased around for fun. The developers watched this happen and pivoted the whole game into it. Slack was a byproduct of a video game that never launched. Minecraft was a fundraiser for a dream game Notch was building, and Tim’s assessment is that Notch made enough money to “build whatever he wants in the physical world, honestly.”
But James dropped the important caveat: every one of those success stories involved massive investment before the feedback started. The racing game already existed. Slack’s network was already built. It wasn’t “throw out garbage and hope.” Tim raised his hand for survivorship bias too — for every glitch-turned-genius, there’s a graveyard of interfaces nobody’s writing nostalgic blog posts about.
The Graveyard Tour
Speaking of graveyards — this is where the episode found its groove. The Google Graveyard and Microsoft Graveyard, those viral catalogs of killed products. Tim cited the stat that roughly half of everything Google has ever built is dead. Google Reader (still mourned), Google Music (the loss that drove Tim to Spotify, because YouTube Music launched without shuffle, a genuine war crime), Google Glass, Chromecast, and Android’s roughly nine hundred attempts at a chat app.
Which teed up Tim’s grand unified theory of technology: everything eventually becomes a chat app. Just as evolution has independently produced crabs four separate times, all software is slowly crawling toward being a chat app. Is chat the crab of the internet? The episode declined to rule it out.
Microsoft’s failures got a more sympathetic read. The Zune was arguably better than the iPod; the Windows Phone’s tiled interface got quietly borrowed by everyone. The problem, per Tim, is that nobody wants to use a Microsoft product for fun. You cannot picture the iPod silhouette-dancing commercial featuring a Zune. It’s just a guy at an office trying to listen to music before the cubicle walls close in.
The Netflix Problem, or: We Reinvented Television
The back half pivoted to TV, and James made the strongest argument of the episode: the “ship eight episodes and cancel if it’s not #1” model has killed the sitcom. Sitcoms are still the most-rewatched shows on earth — people loop The Office forever — but season one of any sitcom is awkward. The writers need time to find the characters; the viewers need time to feel like the characters are friends before the quirks land. Netflix doesn’t grant that time, which is why James is convinced Netflix is structurally incapable of making a sitcom.
Tim’s been rewatching Malcolm in the Middle with his daughter, who cannot comprehend that families used to own one television and fight over it. But he sees the model swinging back: something like 30% of releases are now weekly, because streamers finally noticed that people subscribe for one month, binge two shows, and cancel. Drip-feed the episodes, and you can extract a full year of subscription instead. In other words, they ran the numbers, invented weekly episodic television, and called it innovation. We reinvented TV.
The Part Where You Realize You Own Nothing
The conversation curdled — productively — into ownership. Right-to-repair, John Deere’s tractor software mess, and the Canadian tractor company gaining ground by selling farmers a tractor with no software, which turns out to be exactly what a lot of farmers want. Then the personal betrayals came out. James bought two Gen 1 Nest thermostats, loved them, and watched Google sunset the servers — no open-sourcing, no letting a developer hack away at hardware physically bolted into his house. Just: no. Tim eulogized his original Pebble watch, which was killed when Fitbit (owned by Google — it always comes back to the big companies) shut down the servers.
The thesis crystallized here: once a company sheds you as a user, you don’t come back. Not because their product got worse, but because you’ve got a bin full of abandoned hardware and you remember it every single time you reach for their brand. Tim went from original Android on an HTC to almost entirely Apple, one killed product at a time. Somewhere, someone is running the numbers on that, and they’re not good.
Apple didn’t escape clean, either. James is still bitter about the Mobiblu Cube — a one-inch MP3 cube he adored — which Apple bought, shut down, and cannibalized into the screenless, buttonless Shuffle. And Dark Sky, the beloved weather app, bought and buried. Apple just does it on the down-low, brand unattached.
Bring Back Clippy
The best tangent: Tim’s claim that Clippy was the original ChatGPT. Think about it — Clippy popped up uninvited, going “looks like you’re writing a resume, do you need help?” and then generated a bunch of trash you didn’t ask for. “Is that not what the agents are doing today?” James floated the idea that, amid everyone hating Copilot being shoved into everything, Microsoft’s smartest move would be to just bring Clippy back. Tim agreed — an entire generation would use Clippy out of pure spite and nostalgia, quality be damned. The nightmare version: Clippy on Microsoft Teams, popping out of the corner mid-message, going, “You’re trying to tell your boss you’ll blow this deadline. Do you need help?”
Would You Rather: Corporate Abandonment Edition
Round 1: Only use products from huge companies (guaranteed to end up in a graveyard eventually) — or only from small startups (acquired and shut down eventually)? Tim’s logic: startups only ever end in acquisition or becoming the behemoth, so you’re landing at a big company either way. He leans startup for pure software, but he’s been bitten — a recipe app called ChefTap died because the developer lost his Apple developer key and rage-quit, stranding all of Tim’s recipes. His solution is now to print recipes into a binder, like a colonial homesteader. James admitted all his recipes live in Google Docs, so his ideals and his behavior are, quote, not aligned.
Round 2: First-class support for everything you pay for — or for everything open-source with full manual access? Tim opened with the truth that first-class support does not exist; the best case is someone who solves your problem while being visibly annoyed you called. So he’ll take open source, provided the loophole holds that he can pay a third party to fix it — because he is not taking his fridge apart to check the ice maker’s motherboard. James agreed the instant third-party experts entered the equation.
Round 3: Lots of little competing chat options (AIM, ICQ, MSN, all disconnected) — or one or two big ones (Slack, Teams) but at their mercy? Tim wants the little guys in theory, but remembers that in practice everyone just downloaded Trillian to talk across all of them anyway. You always end up at a unifier. The difference: Slack and Teams have zero incentive to interoperate, so a Slack company and a Teams company will never communicate in anything other than email. This led to the episode’s thesis on email itself — that it only works because it was invented before companies owned the internet, back when nobody pitched a TAM for sending messages. “You couldn’t build email today,” Tim argued. You couldn’t build libraries today either — try pitching free books for anyone to America in 2026.
The Takeaway
Ship fast, sure. Iterate, fine. But every clean success story had real investment behind it before the feedback started, and for every GTA there’s a landfill of killed products and a customer holding a dead thermostat. As James put it, the episode’s actual theme is simpler than any of the frameworks: at the end of the day, the customer suffers. Just build everything open. Except then you’d have to compete on merit, which is hard — because maybe you built an MVP that turns out to be never actually viable.
Give it time to bake. Or don’t, and cancel it after eight episodes. That’s apparently allowed now.
New episodes drop weekly, which — per this very episode — is the correct and newly-reinvented format for episodic content. Tim’s recipes are safe in a binder.